I am currently a Senior Machine Learning Technologist at Ofcom, working within the Online Safety Group on research relating to recommender systems, and more broadly, responsible AI. I obtained my PhD from The University of Manchester in 2024, where I worked within the Centre for AI Fundamentals on physics-informed probabilistic deep learning under the supervision of Dr Mauricio Álvarez. During my PhD, in 2022, I worked as a research intern at Spotify on bandit algorithms for podcast recommendation.
Previously known as Thomas M. McDonald.
Download my CV.
- Recommender systems
- Responsible AI
- Gaussian processes
- Bayesian deep learning
- Physics-informed ML
-
PhD in Computer Science, 2024
The University of Manchester
-
MSc in Data Analytics, 2020
The University of Sheffield
-
BSc (Hons) in Physics, 2018
The University of Sheffield
Publications
Experience
- Currently, I am a senior researcher in the AI team, within the Online Safety Group at Ofcom. My work currently focuses on topics relating to recommender systems and more broadly, responsible AI.
- I spent the summer working within Tech Research at Spotify on a project which leveraged Bayesian filtering and Thompson sampling to provide podcast recommendations which drive long-term user engagement. This work was presented at KDD 2023.
- I have worked as a GTA on a number of different courses within the Faculty of Engineering, and currently assist with postgraduate-level courses focused on machine learning, handling data at scale using Spark and High Performance Computing infrastructure.
-
My role involved employing statistical modelling to forecast national non-commodity cost components and mitigate the level of risk involved in signing energy supply contracts.
-
Implemented seasonal ARIMA forecasting models in Python, with the models routinely returning <1% error on predictions made three months ahead of time.
-
Improved functionality of the VBA gas and electricity price matrices.